In response to a comment of Michael Brunnbauer and a reply by Ed Barkmeyer: (01)
[MB]
>> How about a dictionary without circular definitions ? I do not know of
>> such a thing and I bet it is not because of ignorance on my part.
>
> The purpose of natural language dictionaries is to inform humans who
presumably have some familiarity with the language.
> So, primitive terms are "defined" by providing synonyms and circular
circumlocutions. The idea is that the human reader
> will recognize enough of that verbiage to grasp the intended concept, by
being familiar with the concept itself, presumably in other terms.
Yes, any "dictionary " that includes only symbols (including existing
ontologies) will not be properly "grounded" in a manner that would permit a
computer to actually "understand" the words in a manner analogous to and
functionally comparable to the way humans understand it. The most common
theories of grounding assume that it is the experience of living in the
physical world and observing how words relate to objects and actions, etc.,
that provides a basis for real "understanding". At present, the "primitive"
concepts are grounded only by the understanding of their meaning by the
humans who use those terms in their own lives or computer programs. For a
computer to have comparable autonomous grounding so that it could, for
example, write its own programs that interact properly with the world, would
require both sensor and effector routines, i.e. robotic capability,
integrated with the ontology. I expect that to happen, but *when* depends
largely on how well research programs aimed in that direction are funded.
Given the current trends and fads, even the usual "twenty years" estimates
for AI goals are probably overoptimistic. It was twenty years ago that it
became evident to some of us that a well-structured ontology integrated with
a language understanding program was a necessary step in that direction, but
in the interim the funding for even that baby step has been seriously below
what is needed. (02)
There is a theory of meaning called "procedural semantics" in which the
meaning of a word describing a physical object is the ability to recognize
it when it is seen, and the meaning of an action is both to recognize it
when seen, and to do it (in favorable cases). That is close to how I view
the was an ontology would be grounded. (03)
Interestingly, though current computers are not totally without some
"grounding". Programs that interact with the internet can access some
sensor information (e.g. local temperature or satellite photos and other
pictures) as well as voluminous text that provides serious constraints on
the meanings of the primitive words and concepts. And, of course, robots
that interact via natural language will have some of the necessary routines
- but I think that the current language-robot systems are too specialized to
serve as a starting point for the more general primitives-based foundation
ontology. I think that the routines in "virtual reality" programs that try
to adhere to real-world physics will also have useful subroutines that might
be used for grounding of an ontology. (04)
So, a dictionary (in the form of an ontology) without "circular
definitions" is definitely possible, and will include not just static
symbols, but functional calls to procedural routines that have standardized
sensory and robotic effects. Among other things, when one asks it to
provide a definition of an action, it will be able to show that action in a
video sequence - and know the relation among the objects in the video, so it
can answer clarifying questions. Really cool , but I fear I will not live
long enough to see it. (05)
Pat (06)
Patrick Cassidy
MICRA Inc.
cassidy@xxxxxxxxx
1-908-561-3416 (07)
-----Original Message-----
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Barkmeyer,
Edward J
Sent: Friday, August 02, 2013 1:48 PM
To: [ontolog-forum]
Subject: Re: [ontolog-forum] Context and Inter-annotator agreement (08)
Michael Brunnbauer wrote: (09)
> Or take the sentence "Soldiers are potential murderers". The sense of
> "murderer"
> is modified by "potential" to be something completely different. (010)
Indeed. This is one of the problems with parsing natural language in terms
of 'parts of speech'. "Adjectives" like "fake" and "potential" do not
"narrow" the concept they modify; they pervert it. In that way, "potential
murderer" and "murderer" are different "noun concepts". (011)
> ... <snip>
> How about a dictionary without circular definitions ? I do not know of
> such a thing and I bet it is not because of ignorance on my part. (012)
The purpose of natural language dictionaries is to inform humans who
presumably have some familiarity with the language. So, primitive terms
are "defined" by providing synonyms and circular circumlocutions. The idea
is that the human reader will recognize enough of that verbiage to grasp the
intended concept, by being familiar with the concept itself, presumably in
other terms. (013)
Formal glossaries, including the electronic versions supported by SKOS and
SBVR, have a notion of 'primitive' or 'commonly understood' term, which
means that the term does not appear in the glossary at all, even though it
is used in definitions, or it appears and has no "definition" (it is marked
'primitive' or 'common'). So, yes, "dictionaries" without circular
definitions do exist, but they don't serve quite the same purpose as the
Oxford English Dictionary or La Petite Larousse. (014)
> My opinion is that understanding of natural language is not possible
> without true intelligence. It may also be a bit the other way round. (015)
The understanding of that sentence is not possible without a definition of
"true intelligence"; it is not a 'commonly understood' term... :-) (016)
-Ed (017)
--
Edward J. Barkmeyer Email: edbark@xxxxxxxx
National Institute of Standards & Technology Systems Integration Division,
Engineering Laboratory
100 Bureau Drive, Stop 8260 Work: +1 301-975-3528
Gaithersburg, MD 20899-8260 Mobile: +1 240-672-5800 (018)
"The opinions expressed above do not reflect consensus of NIST, and have
not been reviewed by any Government authority." (019)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (020)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (021)
|